Automatic cortical gyral parcellation using probabilistic atlas and graph cuts

Automatic parcellation of cortical surfaces into sulci or gyri based regions is of great importance in studying the structure and function of the human brain. This paper presents a novel method for automatic parcellation of cortical surfaces into gyri based regions. The method is composed of two major steps: data-driven gyral patch segmentation and model-driven gyral patch labeling. The gyral patch segmentation is achieved by several steps, including sulcal region segmentation, sulcal basin parcellation, gyral crest segments extraction and gyral patch segmentation. The gyral patch labeling is formulated as an energy minimization problem, in which a cortical probabilistic atlas and the curvature information on surfaces are used to define the energy function. The energy function is efficiently solved by the graph cuts method. A unique feature of the proposed method is that it does not require high dimensional spatial normalization on images or surfaces. The method has been successfully applied to cortical surfaces of 15 young healthy brain MR images. Quantitative and qualitative evaluation results demonstrate the validity of the proposed method.